AI Engine · by Anthropic

How Claude chooses which brands to recommend

How does Claude pick brands?
Claude answers from training memory by default and searches the web when the question needs current information. It is notably conservative about unverified claims — where evidence about a brand is thin or conflicting, it will hedge or decline to recommend rather than guess, which makes corroboration unusually important.
Where it gets sources
Web search when enabled; training memory otherwise
How it attributes
Inline citations when search is used

How it picks who to name

It weights consistency heavily. Conflicting information about a brand tends to produce hedged language rather than a confident recommendation.

When searching, it prefers sources that state things plainly and can be attributed, and will say when it could not find something.

Its caution cuts both ways: thin evidence suppresses you, but clean, consistent evidence is rewarded with unusually direct statements.

How to get cited by Claude

  1. 1

    Allow ClaudeBot and Claude-Web

    ClaudeBot crawls broadly; Claude-Web fetches pages for live questions. Both need to reach you.

  2. 2

    Remove contradictions first

    Because Claude hedges on conflicting evidence, resolving inconsistent pricing and descriptions often produces a bigger lift here than new content.

  3. 3

    Make claims verifiable

    Specific, checkable statements backed by an independent source survive its caution; marketing adjectives do not.

What makes Claude different

The engine most likely to explicitly state that it could not verify something about a brand. That makes it a useful diagnostic — if Claude says it cannot find reviews for you, that gap is real and every engine is seeing it.

Does Claude name you?

We ask real buyer questions and show you the answer, the brands named instead, and every source cited.

Related questions

Other AI engines

Frequently asked about Claude

How does Claude decide which brands to recommend?+

Claude answers from training memory by default and searches the web when the question needs current information. It is notably conservative about unverified claims — where evidence about a brand is thin or conflicting, it will hedge or decline to recommend rather than guess, which makes corroboration unusually important.

How do I get cited by Claude?+

Allow ClaudeBot and Claude-Web: ClaudeBot crawls broadly; Claude-Web fetches pages for live questions. Both need to reach you. Remove contradictions first: Because Claude hedges on conflicting evidence, resolving inconsistent pricing and descriptions often produces a bigger lift here than new content. Make claims verifiable: Specific, checkable statements backed by an independent source survive its caution; marketing adjectives do not.

Where does Claude get its information?+

Web search when enabled; training memory otherwise. Citations: inline citations when search is used.

What makes Claude different from other AI engines?+

The engine most likely to explicitly state that it could not verify something about a brand. That makes it a useful diagnostic — if Claude says it cannot find reviews for you, that gap is real and every engine is seeing it.